Background

We have shown that using age-adjusted admission systolic blood pressure (SBP) values (z-scores) may offer advantages in determining the associated mortality risk in a large unselected PICU population.

Objectives

We hypothesized that the SBP z-score risk profile would differ in children admitted with a primary cardiac diagnosis from the rest of the cohort.

Methods

This is a retrospective cohort study using data from 2 cardiac and 2 general PICUs from 2004-2018. We derived SBP z-scores according to the NIH Task Force definitions of normal SBP (1,2). We defined the relationship between mortality and SBP z-scores using logistic regression, adjusted for PIM-2 with the blood pressure component removed.

Results

Data from 34,745 patients were analysed (19,649 cardiac and 15,096 non-cardiac admissions). The relationship between mortality and SBP z-scores are shown in Figure 1. The relationship is U-shaped for general ICU admission, but inverse for cardiac ICU.

Conclusion

Our data show an inverse relationship between SBP z-scores and mortality in the cardiac population. The decreased risk of mortality with hypertension is potentially representative of the favourable outcome in children with good post-operative recovery of heart function. Whether manipulating blood pressure modifies the risk of death needs to be assessed in an interventional trial.

References:

(1) National Institutes of Health. The fourth report on the diagnosis, evaluation, and treatment of high blood pressure in children and adolescents. Available from:https://www.nhlbi.nih.gov/files/docs/resources/heart/hbp_ped.pdf [Accessed 18/6/2018].

(2) National Institutes of Health-Task Force on Blood Pressure Control in Children. Report of the Second Task Force on Blood Pressure Control in Children-1987.Pediatrics. 1987. 79;1:1-25.

Background

Mortality is modelled as a quadratic function of systolic blood pressure (SBP) at admission to PICU in the Paediatric Index of Mortality score (PIM). The lowest risk of mortality is at 120 mmHg, regardless of age (1). As most children on PICU, are under 2 years of age, this is the >99^th centile.

Objectives

We sought to define the association between age-corrected SBP z-scores and mortality in critically ill children.

Methods

In this retrospective cohort study, we gathered SBP values on admission to four paediatric intensive care units (2 cardiac and 2 general) in 34,745 children from 2004-2018. We derived SBP z-scores according to the NIH Task Force definitions of normal SBP. We defined the relationship between mortality and SBP z-scores using logistic regression, adjusted for PIM-2 with the blood pressure component removed.

Results

The relationship in our cohort showed a U-shaped curve with a nadir at SBP between 111-120, or 2.25<=z<2.75 (Fig.1), though mortality was found to be relatively invariant over a wide range of SBP values (increasing only from bands 1.25<=z<1.75 and >=3.25) with narrow confidence intervals.

Conclusion

Our data support the value of using SBP z-scores rather than absolute values in children. Prospective trials examining the impact of different blood pressure targets on mortality in critically ill children are needed.

References:

(1) Shann F, Pearson G, Slater A, Wilkinson k. Paediatric index of mortality (PIM): a mortality prediction model for children in intensive care. ICM. 1997. 23:201-207.

Background

The relationship between temperature and heart rate (HR) has been described as non-linear, with HR increasing by ~ 10 beats/min for every 1°C rise in body temperature. Previous studies have excluded critically ill children.

Objectives

We sought to delineate this relationship in the PICU population using high-frequency bedside monitoring data.

Methods

This is a retrospective observational study using the Etiometry T3 recording system of 142 children with continuous central temperature monitoring admitted to the PICU/NICU at Great Ormond Street Hospital in 2016. HR data were standardised to provide z-scores. Multi-level linear regression analysis of 7,393,328 HR measurements was done for 1°C intervals between 36-40°C.

Results

The regression coefficient for temperature was 8.314 (95% CI 8.299-8.329). The relationship between temperature and HR was non-linear, with the greatest change between 38-39°C, with HR increasing by 20 beats/min (Figure 1). The relationship remained unchanged when analysis was repeated using age-standardised HR.

Conclusion

The relationship between HR and temperature is non-linear in critically ill children. Fever grade (≥38°C) temperature is associated with a significant increase in HR. Further studies are needed to understand the impact of the raised heart rate on haemodynamics in critically ill children

Background

We have shown that using age-adjusted admission systolic blood pressure (SBP) values (z-scores) may offer advantages in determining the associated mortality risk in a large unselected PICU population.

Objectives

We hypothesized that the SBP z-score risk profile would differ in children admitted with a primary cardiac diagnosis from the rest of the cohort.

Methods

This is a retrospective cohort study using data from 2 cardiac and 2 general PICUs from 2004-2018. We derived SBP z-scores according to the NIH Task Force definitions of normal SBP (1,2). We defined the relationship between mortality and SBP z-scores using logistic regression, adjusted for PIM-2 with the blood pressure component removed.

Results

Data from 34,745 patients were analysed (19,649 cardiac and 15,096 non-cardiac admissions). The relationship between mortality and SBP z-scores are shown in Figure 1. The relationship is U-shaped for general ICU admission, but inverse for cardiac ICU.

Conclusion

Our data show an inverse relationship between SBP z-scores and mortality in the cardiac population. The decreased risk of mortality with hypertension is potentially representative of the favourable outcome in children with good post-operative recovery of heart function. Whether manipulating blood pressure modifies the risk of death needs to be assessed in an interventional trial.

References:

(1) National Institutes of Health. The fourth report on the diagnosis, evaluation, and treatment of high blood pressure in children and adolescents. Available from:https://www.nhlbi.nih.gov/files/docs/resources/heart/hbp_ped.pdf [Accessed 18/6/2018].

(2) National Institutes of Health-Task Force on Blood Pressure Control in Children. Report of the Second Task Force on Blood Pressure Control in Children-1987.Pediatrics. 1987. 79;1:1-25.

Background

Mortality is modelled as a quadratic function of systolic blood pressure (SBP) at admission to PICU in the Paediatric Index of Mortality score (PIM). The lowest risk of mortality is at 120 mmHg, regardless of age (1). As most children on PICU, are under 2 years of age, this is the >99^th centile.

Objectives

We sought to define the association between age-corrected SBP z-scores and mortality in critically ill children.

Methods

In this retrospective cohort study, we gathered SBP values on admission to four paediatric intensive care units (2 cardiac and 2 general) in 34,745 children from 2004-2018. We derived SBP z-scores according to the NIH Task Force definitions of normal SBP. We defined the relationship between mortality and SBP z-scores using logistic regression, adjusted for PIM-2 with the blood pressure component removed.

Results

The relationship in our cohort showed a U-shaped curve with a nadir at SBP between 111-120, or 2.25<=z<2.75 (Fig.1), though mortality was found to be relatively invariant over a wide range of SBP values (increasing only from bands 1.25<=z<1.75 and >=3.25) with narrow confidence intervals.

Conclusion

Our data support the value of using SBP z-scores rather than absolute values in children. Prospective trials examining the impact of different blood pressure targets on mortality in critically ill children are needed.

References:

(1) Shann F, Pearson G, Slater A, Wilkinson k. Paediatric index of mortality (PIM): a mortality prediction model for children in intensive care. ICM. 1997. 23:201-207.

Background

The relationship between temperature and heart rate (HR) has been described as non-linear, with HR increasing by ~ 10 beats/min for every 1°C rise in body temperature. Previous studies have excluded critically ill children.

Objectives

We sought to delineate this relationship in the PICU population using high-frequency bedside monitoring data.

Methods

This is a retrospective observational study using the Etiometry T3 recording system of 142 children with continuous central temperature monitoring admitted to the PICU/NICU at Great Ormond Street Hospital in 2016. HR data were standardised to provide z-scores. Multi-level linear regression analysis of 7,393,328 HR measurements was done for 1°C intervals between 36-40°C.

Results

The regression coefficient for temperature was 8.314 (95% CI 8.299-8.329). The relationship between temperature and HR was non-linear, with the greatest change between 38-39°C, with HR increasing by 20 beats/min (Figure 1). The relationship remained unchanged when analysis was repeated using age-standardised HR.

Conclusion

The relationship between HR and temperature is non-linear in critically ill children. Fever grade (≥38°C) temperature is associated with a significant increase in HR. Further studies are needed to understand the impact of the raised heart rate on haemodynamics in critically ill children